1

The Association of Compact Groups of Galaxies with Large-scale Structures

Heinz Andernach & Roger CoziolDepartamento de Astronomía

Universidad de Guanajuato, Mexico

2

• Groups of galaxies = principal environment of mostgalaxies in the Universe, but …their origin & evolution is still poorly understood(e.g. resolution of simulations is still insufficient)

• Important questions (Hierarchical Structure Formation):– When do they form? Suppose biased galaxy formation

is there a “downsizing” effect for groups ?(high-mass groups at high z form first, then lower-mass groups at lower z)

– What is the role of groups in the formation of larger structures? How long can a group retain its dynamical characteristics when merging with other groups to form larger structures?

– What is the importance of groups for the formation and evolution of galaxies in large-scale structures?

3

• Observationally, the intrinsic low richness of groups makes them moredifficult to detect than richer systems like cluster of galaxies, especially at intermediate and high redshift.

• Easier to detect: Compact Groups of Galaxies (CGs)1. Small mass systems (formed of 4 to 8 galaxies) 2. High spatial overdensities (comparable to clusters)3. Low velocity dispersions (<<1000 km/s)

• Recent efforts to form “objectively” selected samples of CGs at low and intermediate redshifts (e.g. SCGs, SDSS-CGs and PCGs)

Goals of our study: relation of CGs with large scale structures

– Differences between selection algorithms,e.g. eye vs automatic

– The nature of association of CGs with large scale structures: flukes in selection method or real dynamical “substructures”?

– Variation with redshift: observational biases or evolution?

4

Previous Studies:

• Rood & Struble (1994): enviro of HCGs (low redshift, z ~ 0.04)out to about 2 Abell radii:

- some HCGs are associated with Abell rich clusters, and - many more (~2/3) with poorer systems (“Loose Groups”)

Although CGs are “isolated dynamical systems” their formation must be related to the formation of the large-scale structure

• It is not clear if CGs associated with large-scale structures form real dynamical substructures or transient compact configurations(Diaferio, Geller & Ramella 1994, Governato, Tozzi & Cavaliere 1996; Mamon 1986; Walke & Mamon 1989)

5

Different studies on the association of CGs with larger-scale structures at low redshift suggest that

groups form after the formation of larger structures: 1. West (1989): clusters of galaxies exhibit a tendency

to align with their neighbors on scales ~15-30 h-1 Mpc massive structures formed before low-mass ones

2. Einasto et al. (2003): Loose Groups in the neighborhood of rich clusters seem more massive and luminous than groups on average their formation process depends on those of the more massive systems?

3. Coziol, Brinks & Bravo Alfaro (2004): galaxies in CGsassociated with massive structures seem more evolved, in terms of morphology and level of activity, than those in CGs associated with lower-mass structures

difference in time of formation (global effect) ordifference in the rate of galaxy evolution (local effect)?

6

Samples used in our study

• Compact groups:

- 100 Hickson Compact Groups, HCGs (Hickson 1982)- 121 Southern Compact Groups, SCGs (Iovino 2002)- 459 Palomar Compact Groups, PCGs (deCarvalho et al. 2005)- 177 SDSS Compact Groups, SDSS-CGs (Lee et al. 2004)

• Associated Large-Scale Structure:- Abell clusters: 5250 clusters (Abell/ACO + suppl. S-clusters;

most recent compilation of z’s (Andernach & Tago, >3000 z’s)- Northern Sky Optical Cluster Survey Catalog, NSC

(Gal et al. 2003); only poor photometric redshifts

- Loose groups: UZC-SSRS2 group catalog (Ramella et al. 2002);all with known (low) redshifts

7

Comparison Abell vs NSC clustersNSC clusters not associated with Abell clusters are poorer:

Number of galaxies

Ngal(NSC)(non-assoc)

Ngal(NSC)(Abell-assoc)

mean 29 43median 30 41

NSC reach slightly higher redshifts than the Abell sample: Redshifts NSC Abell

mean 0.1616 0.1541median 0.1565 0.1468

NSC does not include smaller structures similar to loose groups:incompleteness at redshift z ~ 0.1 could be as high as 70% according to NSC authors (Gal et al. 2003).

8

Method• Association with Abell clusters: if projected distance from

CG to ACO center is <1 Abell radius of the cluster (Abell radius = 1.5 h-1 Mpc, with H 0 = 100 h-1 km s-1 Mpc-1 );

• Association with NSC and UZC-SSRS2: cross-correlation of thelatter catalogs with CG catalogs available at Strasbourg Data Center (CDS, Vizier) within one Abell radius.

• Confirming the associations:a) Visual inspection on DSS2 images from Aladin (aladin.u-strasbg.fr)

b) Comparing redshifts:HCGs (Hickson et al 1992) and SDSS-CG (Lee et al. 2004);SCGs, mean value of galaxies included in NED;

? PCGs, very few redshifts available (finding charts were not published);

9

Type of associationFrom our visual inspection, we classify the type of association thatCGs may have with the larger systems; depending on

a) projected distance of the group relative to the center of its associated structure and

b) relative apparent magnitudes of the galaxiesin the two structures

we distinguish three classes of apparent association:

1) SS: CGs are not cluster-centric; seem to form “substructure”; CG galaxies are either fainter or as luminous as the brightest galaxies in the larger system;

2) ML: CG forms main structure itself (centric position);CG members are the “most luminous” of entire system

3) AML: CG members are part of the main structure;galaxies are “among the most luminous” of the system

10

• Include here examplesHCG 65 = A3559 (ML type)

R = 3 BM = I

11

HCG 63 (SS type)v = 9324 km/s

A3537, R=0, BM = I-IIv = 5100 and 9593 km/s

12

PCG J100102-001342 (AML type)

NSC J100116-001126

13

Association with Abell clusters

Difference in method: - HCG vs SCG: eye selection reduces number of CG-LSS

associations: 83% SS type in HCGs vs 57% ML in SCGs.- SDSS vs PCG: 69% SS type in SDSS vs 72% AML in PCGs

(why only 1 SDSS-CG coincides with a PCG ? Expect ~10)

- Association rate is similar at low & intermediate z, BUT :

a) Richness of associated clusters rises with z:fraction of R ≥ 1 cluster rises: 14% SCGs to 69% (PCGs & SDSS)

b) BM type changes with z: from 74% early (I, I-II and II) in SCGsto 92% late (II-III, III) in the PCG and 100% “late” in the SDSS.

14

Variation of Abell cluster richness with zFrom z = .02 to .1 the frequencyof poor ACO clusters with R=0decreases only by factor of 1.3.

… while CG-associated R=0 ACOclusters decrease by factor of 6passing from SCGs to PCGs.

Fraction of ACO clusters with R >=1 rises only by a factor of 2.

… while CG-associated ones riseby factor of 5 from SCGs to PCGs

Increase in richness of the PCG-associated ACO-clustersat intermediate z cannot beexplained by an incompletenessof the Abell/ACO sample.

15

Variation of BM types with z

ACO sample @ z~0.1composed of 48/52 %

early/late BMs.~ 50-50 chance to

find early or late BMtype at this redshift.

Indeed we find not one CG-associated ACO cluster of BM I or I-II @ z~.1.This is despite the equal detection probabilities for early & late BM types.

preference of late BM types for CG-associatedclusters at intermediate z is not due to a selection effect in the ACO sample.

16

Association of CGs with poorer structureAssociations with UZC-SSRS2:

Associations with NSC:

Rate of associations risesw.r.t. cluster-associations:

factor 6 for HCG and 2 for SCG

factor 4 for PCG and 3 for SDSS

Given the incompleteness,most CGs could be associatedwith larger structures atintermediate redshift.

Majority are ML or AML typeCGs form importantcomponents of structures.

17

Possible interpretationVariations with redshift of Abell richness, association type and BM type are consistent with an increase of mass of the associated structures at intermediate redshift.

No obvious evidence of observational bias for richness or BM type(no sufficient decrease of poorer ACO or NSC cluster densityseen out to z = 0.1;

early/late BM types = 50/50 at z = 0.1 in ACO)

Same trend is observed despite the differences introduced by theselection methods

general phenomenon ?

Association types ML and AML more frequent at intermediate z and in poorer structures

CGs may generally form an important “substructure” of the associated larger-scale structure.

18

Evolution of galaxies in CGs: morphological type distributionTypes from CDS catalog VII/213 Types from NED (brightest ~37% galaxies)

• Galaxies in CGs associated with larger scale structures are moreevolved morphologically associations are physically real

• High number of early-type in isolated CGs CG favor evolution• No difference between isolated and loose groups

threshold in density or difference in formation time?

HCG SCG

19

Cautionary notes on automated (compact) group detection

While inspecting DSS images of PCGs we noticed “doubtful” caseswe inspected ALL 459 PCG images of DSS (not contained

in deCarvalho et al. 2005, contrary to author’s statement)

We found 52/459 (11%) of PCGs likely contaminated withstellar images (MANY confirmed by negative g-r colors aslisted by authors themselves in their table of members)

SDSS-CGs have very different morphology than PCGs, many look like single galaxies with rather “insignificant”satellites Do they represent the same type of objects?

The 2dF “group” catalogue (V. Eke et al. 2004) contains severaldozen rich Abell clusters (Nmem >~ 100)

PCG J164047+324843B and C starlike

Cluster NSC J164051+324856at left edge of image

PCG J154942+173434A,B,C are starlike with

negative colors

20

A

BC

DA

BC

DA

B D

C

PCG J1617+2041B+D = part of MCG +04-38-044(v = 9027 km/s)A and C starlike

PCG J1618+4028B (extreme color) right on a

diffraction streak of star

21

B

CD

A

A

BC

D

22

ConclusionsWe should return to a clear terminology differentiating betweenpoor groups, compact groups, loose groups, clusters, rich clusters…(or else use terms like “overdensities”)

Visual inspection of automatically “detected” groups is mandatory.

CG could play an important role in the formation and evolution oflarge scale structures and in the evolution of galaxies in thesestructures. This would be consistent with the actual paradigm ofHierarchical Structure Formation.

The variation in R and BM-type we observe from to medium zmay also be consistent with biased galaxy formation.

For other evidence in favor of this model see two posters:- Kinematics of BCM (Andernach et al.) and on- Evolution of galaxies in CG (Plauchu et al.)

23

Job vacancy: tenure track position in MexicoA view from our premises onto colonial Guanajuato

visit us at http://www.astro.ugto.mx

We are located at the geometric center of Mexico:

24

We seek: extragalactic astronomer/observational cosmologistContact me/us at heinz@astro.ugto.mx

25

Variation of Abell cluster richness with zFrom z = .02 to .1 the frequencyof poor ACO clusters with R=0decreases only by factor of 1.3.

… while CG-associated R=0 ACOclusters decrease by factor of 6passing from SCGs to PCGs.

Fraction of ACO clusters with R >=1 rises only by a factor of 2.

… while CG-associated ones riseby factor of 5 from SCGs to PCGs

Increase in richness of the PCG-associated ACO-clustersat intermediate z cannot beexplained by an incompletenessof the Abell/ACO sample.

26

Variation of NSC abundance with redshift• The phenomenon we observe for the CGs is consistent is with a genuine decrease of

small-scale structures at higher redshift and, consequently, of an increase of structure masses at higher redshift.